Test failure while building cpython

Ned Deily nad at acm.org
Wed Aug 20 20:25:05 EDT 2014


In article <09843563-B0FD-451B-BF66-0FB720CEC67A at ja.net>,
 Adam Bishop <Adam.Bishop at ja.net> wrote:
> I'm trying to build python 3.3.2 from source packages provided by Red Hat 
> under mock.
> 
> The build itself works,but one specific test is failing:
>   
>    ======================================================================
>    FAIL: test_sysconfig_module 
>    (distutils.tests.test_sysconfig.SysconfigTestCase)
>    ----------------------------------------------------------------------
>    Traceback (most recent call last):
>      File 
>      "/builddir/build/BUILD/Python-3.3.2/Lib/distutils/tests/test_sysconfig.py
>      ", line 134, in test_sysconfig_module
>        self.assertEqual(global_sysconfig.get_config_var('CFLAGS'), 
>        sysconfig.get_config_var('CFLAGS'))
>    AssertionError: '-Wno-unused-result -DDYNAMIC_ANNOTATIONS_ENABLED=1 -O2 -g 
>    -pipe -Wall -Wp,-D_FO [truncated]... != '-Wno-unused-result 
>    -DDYNAMIC_ANNOTATIONS_ENABLED=1 -O2 -g -pipe -Wall -Wp,-D_FO 
>    [truncated]...
>    - -Wno-unused-result -DDYNAMIC_ANNOTATIONS_ENABLED=1 -O2 -g -pipe -Wall 
>    -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
>    --param=ssp-buffer-size=4 -grecord-gcc-switches   -m64 -mtune=generic 
>    -D_GNU_SOURCE -fPIC -fwrapv -I/usr/include-O2 -g -pipe -Wall 
>    -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
>    --param=ssp-buffer-size=4 -grecord-gcc-switches   -m64 -mtune=generic 
>    -D_GNU_SOURCE -fPIC -fwrapv  -I/usr/include-O2 -g -pipe -Wall 
>    -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
>    --param=ssp-buffer-size=4 -grecord-gcc-switches   -m64 -mtune=generic 
>    -D_GNU_SOURCE -fPIC -fwrapv   -O1
>    ?                                                                          
>                                                                               
>                                                                               
>                                                                               
>                                                                               
>                                                     
>    ---------------------------------------------------------------------------
>    ---------------------------------------------------------------------------
>    --------------------------------------------------
>    + -Wno-unused-result -DDYNAMIC_ANNOTATIONS_ENABLED=1 -O2 -g -pipe -Wall 
>    -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
>    --param=ssp-buffer-size=4 -grecord-gcc-switches   -m64 -mtune=generic 
>    -D_GNU_SOURCE -fPIC -fwrapv -I/usr/include-O2 -g -pipe -Wall 
>    -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector-strong 
>    --param=ssp-buffer-size=4 -grecord-gcc-switches   -m64 -mtune=generic 
>    -D_GNU_SOURCE -fPIC -fwrapv
>    
>    
>    ----------------------------------------------------------------------
>    Ran 189 tests in 1.778s
> 
> Obviously I could simply disable the test[1], but this requires patching the 
> source package, which I'd rather avoid because I don't fully understand what 
> case this test is trying to detect.
> 
> Is anyone able to indicate what the test is looking for and what might cause 
> it to fail? I don't see how the CFLAGS used in the build process could change 
> during the test run.

The test is to check that the two separate implementations of sysconfig 
in the standard library (don't ask why!), the newer sysconfig and the 
legacy distutils.sysconfig, return the same results for certain key 
configuration variables, like CFLAGS.  It's not the most critical test 
in the world.  The important point is that the values from the 
distutils.sysconfig version are what will be used during C extension 
module builds when you install third-party packages and, if necessary, 
most of them can be overridden, either in the package's setup.py or by 
setting environment variables.  Also, be aware that Python 3.3.2 is 
obsolete.

-- 
 Ned Deily,
 nad at acm.org




More information about the Python-list mailing list